How Modern ATS Systems Score Resumes: A Technical Deep Dive

Introduction

Applicant tracking systems (ATS) have become the gatekeepers of hiring processes in many companies. But how do these systems actually score resumes? In this article, we'll take a closer look at the technical aspects of modern ATS systems, specifically Workday, Greenhouse, Lever, and iCIMS.

Parsing: What's Extracted

When you upload your resume to an ATS system, it undergoes parsing - a process that extracts relevant information from the document. The extracted data includes:

Contact information (name, email, phone number)
Work experience (job title, company, dates of employment)
Education (degrees, institutions, graduation dates)
Skills (keywords, certifications)

The quality of parsing depends on the resume format and content. PDFs are notoriously difficult to parse due to their complex layout and font rendering. Some ATS systems use Optical Character Recognition (OCR) techniques to extract text from images, but this can lead to errors and inaccuracies.

Scoring: Position-on-Page Weighting, Keyword Density, Semantic Match

Once the resume data is extracted, the ATS system assigns a score based on various factors:

Position-on-page weighting: Keywords and phrases found near the top of the page are given more weight than those buried deeper.
Keyword density: The frequency of relevant keywords within the resume text affects the overall score.
Semantic match: ATS systems use natural language processing (NLP) techniques to analyze the semantic meaning behind words, rather than just their literal meaning.

These scoring mechanisms are based on various algorithms and models, but they're not foolproof. For example, a study by Glassdoor found that 58% of resumes were rejected due to formatting issues or missing keywords.

Common Reasons for Auto-Rejection

ATS systems can be finicky, and even the most well-structured resume can fall victim to auto-rejection. Here are some common reasons why:

PDF-as-image: ATS systems struggle to parse PDFs that contain images of text.
Fancy fonts: Uncommon or decorative fonts can confuse OCR algorithms.
Two-column layouts: Resumes with multiple columns can be difficult for ATS systems to navigate.
Missing JD keywords: If a resume doesn't include relevant keywords from the job description, it may not pass muster.
Keyword stuffing detection: ATS systems are designed to detect and penalize resumes that excessively use keywords.

What Actually Fixes a Low-Match Resume

While some advice on improving ATS scores is misguided (e.g., "just use a one-column template"), there are evidence-based strategies that can help:

Re-ordering by JD priority: Prioritize the most relevant information based on the job description.
Semantic rewriting: Use language processing techniques to rephrase and refine your resume content.
DOCX export over PDF: Switch from PDF to DOCX format, which is more easily parsed by ATS systems.

Conclusion

ATS systems are complex beasts that require a nuanced understanding of their inner workings. By grasping the technical aspects of modern ATS systems, you can improve your chances of getting noticed by hiring managers and recruiters.

If you want to try this for yourself, ANANTA Trade offers an autonomous AI system that can help with resume parsing and scoring. You can access a demo version at <https://app.anantatrade.com/?demo=1>.